Stephan G. Kohlmann
Some smart professor once stated: Communication occurs when the right person says the right things in the right way to the right people, and the message is heard and understood. Well, taking this concept and expanding it to include our four legged hunting pals, we could restate the professor’s wisdom to a simple dog trainer’s credo: Dog training occurs when a handler does the right thing in the right moment, and the command is heard and enforced at all times. This principle is easy to subscribe to, but very difficult to adhere to, partly because we have to change our way of communication.
Over the years, I have noted that most people who brought their dogs to me for training had not fully understood this concept, and as a consequence made mistakes which essentially prevented their dogs from learning the right thing. When one analyzes human-canine communication one should first consider the abilities of both parties in verbal and cognitive regards.
THE WORLD ACCORDING TO FIDO
It is no great mystery that dogs are NOT humans without speech. They lack more than a suitable larynx to pronounce their desires in correct English. But they are not dumb brutes, either. Dogs are highly social animals, and as such are very effective communicators in those areas where their survival and genetic fitness is concerned. These areas are naturally food finding, reproduction, and avoidance of physical harm. Unfortunately, the word “sit” has very little to do with any of these natural concerns for our dogs. On the contrary, our dogs do lots of things that are in direct connection with their primeval survival instinct, such as chase, get into trash, attack, roam, hide, etc. In other words: digging up your flower bed to catch a gopher has a lot more to do with a dog’s instincts than nicely sitting at the edge and “admiring” your flowers. In addition, dogs are very good observers, and they are experts when it comes to relating two events or situations together. A Russian behaviorist by the name of Pavlov showed this impressive ability by ringing a bell every time a dog was about to be fed. After a short while, the dog started salivating on hearing the bell alone, without seeing or smelling any food at all. This concept later was termed “conditioned response”, because the dog had to go through some conditioning before the response could be elicited by an apparently unrelated stimulus.
In many ways people are very similar to dogs, and children’s psychologists will tell you that much of human learning appears to be nothing but conditioned responses. So here we have a common link, that should be profitable when training your dog. On the other hand, most people have the ability to also deduct certain causal relationships, and apply learned knowledge in completely novel situation. This is where our dog’s don’t do very well. For example, a child will experience that lying on a soft mattress in the sun is nice. However, if the sun moves, and the mattress is in the shade, the child will soon figure out how to move the mattress into the sun again. Very few dogs master that degree of “intelligence”.
RELATING BEHAVIOR WITH REWARDS
Having established the fact that dogs learn best when we involve conditioned responses, we now have to ask: how can I elicit the desired response on a verbal or visual signal? Remember, the command has no natural meaning — the dog has to “learn” to relate a command (visual or verbal signal) with a desired state of its own being (food, sex, or absence of harm).
In training a dog, we now simply need to figure out how to cause the dog to relate a signal with a desired reward. Coming back to the three pillars of dog motivation (Food, sex and avoidance of harm), we could now construct situation where we give a certain command or signal when the dog experiences one of the three “motivators”.
For example, you can feed your dog tidbits every time he comes when called, but never when he comes without being called. Soon the dog will run up to you for a treat when you call. This works well in most day-to-day situations, and is in fact used by any successful trainer. Pats on the head or praise fulfill the same purpose of making the dog feel good (reward for following a signal). Keeping in mind that the dog is after the reward when he follows the “command”, and by no means is he interested in “pleasing you”, it is easily understood why this method of training is often referred to as “reward-based” training.
However, reward-based training has its major drawback when it comes to stimuli or situations that promise the dog greater rewards than can be obtained by following the signal. Dogs and many higher animals (including our own species) continually strive for the highest possible rewards, a necessary consequence of natural selection. Individuals that adopt behaviors that provide them with more food, sex, and less injury or harm will produce more offspring, which in return will inherit these traits. This process is nothing but a natural arms race: who can get more success in garnering resources?
Coming back to our reward-based trained dog, we can quickly see, where the dog’s inclination to go after the currently highest reward will compromise our training program: the cat (possible food) crossing the street is much more of an incentive than the “attaboy” we have to offer! Now we have to invent a situation where we no longer rely on positive reinforcements (i.e. rewards) but on negative sensations, that will reduce the rewards obtained by chasing the cat.
REDUCING REWARDS
This is where most people fail, because rather than providing a positive reward (food, attaboy) we must instantaneously apply a negative reinforcement, that the dog will link with the stimulus (cat). If it is strong enough, it makes the dog re-examine its choice to chase the cat. The problem here also has to do with the “right way” and the “right time” of applying this negative reinforcement. Few soft and sensitive dogs are susceptible enough to sense their owners state of mind, and will simply avoid causing “anger” in their owners. This is really nothing else than avoiding “harm”, because the dog equates human “angry behavior” with social aggression, just like the puppy learns to get out of the way of the grumpy old male in the pack. But for most dogs, simply yelling at old Fido while he chases the cat isn’t enough of a reduction in the reward he get’s out of the chase. After all, you don’t make the cat run / disappear any faster or less fun to chase!
Obviously, since the dog is going for the highest reward, you must compete with the kick he gets out of chasing the kitty. No activity that comes either before or after the cat is visible will accomplish the reward reduction. The critical moment is during the chase, and all other measures before or after the chase are useless. That is why dogs who are “punished” after a 3-day stray (i.e., a fun trip filled with sex and adventure) will NOT stay home the next time around. By beating the dog upon return, the owner doesn’t realize that the dog already had experienced the rewards for straying, and that the punishment only serves to reduce the current rewards of coming home (food, a warm place to sleep, and human companionship). The consequence? Fido will stay out longer, and he will slink home, hoping nobody will notice his return. No sign of a “bad conscience”, just fear!
If we consider the dog’s desire to maximize his rewards, we can turn dog training into a positive situation for both the trainer and the dog. In this training program, the dog is “shown” a way to maximize its rewards, which will cause the dog to be highly motivated. Wagging tails will almost certainly result!
NO PAIN NO GAIN
Having established that the timing of a negative reinforcement is critical (i.e. we now know the “right time”), we need to think about the “right way” of applying the negative reinforcement. Obviously, only one of the three pillars of dog motivation (food, sex, avoidance of harm) is associated with a negative reward, and that is “harm”. One could also think about withholding food or sex, but this would be difficult to administer and control in a timely fashion. The only motivator that will override any and all others is pain!
In an evolutionary context, the animal that engages in an activity promising future rewards, while neglecting painful consequences will not leave many offspring! Just think of the coyote trying to subdue a porcupine. Most likely he will not survive to reap the rewards (protein) from the behavior (attacking). The pain is an early and usually heeded warning sign that the rewards may not be worth the cost. Only humans continue running a race when their bodies tell them they’re about to incur grave injury.
Pain or bodily discomfort is usually a sure-fire way to reduce rewards associated with behavior the handler deems undesirable. Flipping a chasing dog through the air via a check chord is one way to reduce the rewards of the chase — after a few episodes the dog will remember that flips always follow a chase (or ignoring a signal). The chase looses its reward, and at the same time coming back to the handler offers the highest reward possible (attaboy).
The handler’s art consists simple of applying a negative stimulus (pain) while the dog is experiencing rewards from an undesired activity. However, unless one is willing to apply enormous level of pain, a single repetition is usually not sufficient. Repeated application is necessary to reinforce the response. To be effective, the negative stimulus must come at exactly the same moment each time the dog engages in the activity.
Even we humans become much quicker convinced the less “variance” we experience in certain causal conditions. If I burn my hand every time I try to retrieve a hot muffin from the oven, I usually learn fast to avoid the pain, regardless how tasty the muffin. However, it may take me much longer to figure it out if I get burned only 10% of the time and get to eat a delicious muffin right away the other 90% . The same applies to your dog. If every disregard of your signal is followed by a painful or discomfortable experience, he will be obedient in a very short time, provided there are other rewards available (attaboys). But if he “gets away with it” (i.e. reaps rewards for the chase, etc), he will never completely give up “testing” the situation. The more predictable the consequences, the faster the dog learns to avoid them!
(In the next newsletter, I will expand on this idea and share some of my experience with using electric collars efficiently)
Comentários